OBJECTIVE: Tocilizumab (TCZ) is the only biologic approved in Giant Cell Arteritis (GCA). In clinical trials around a quarter of patients relapse during TCZ treatment. We assess the frequency, features and factors associated with relapses in a wide series of GCA patients in a real-world setting. METHODS: National multicenter observational study of GCA patients treated with TCZ between 2016 and 2021. The variables collected at TCZ initiation were demographic, clinical, laboratory, temporal artery biopsy, and imaging findings, corticosteroids dose, previous therapies and TCZ therapeutic schedule. We perform a comparative study between patients with/ without relapses (bivariate analysis) and a study of factors associated with relapse (multivariate logistic). RESULTS: We study 407 patients (295 women
mean age 73.6 ± 8.9 years). After a mean follow-up of 25.3 ± 21.7 months, relapses were observed in 63 of 407 (15.5 %) patients. At TCZ initiation, no differences were observed between both groups (with/without relapses) in demographic, clinical and laboratory features or corticosteroid dose. The median time to the first relapse was 12 [6-24] months being the most frequent manifestations polymyalgia rheumatica (47.6 %), and headache (12.7 %). In multivariate logistic regression analysis, the set of variables associated with GCA relapses were TCZ initiation later than 6 weeks (OR 3.446 [1.196- 9.931]), optimization (OR 2.803 [1.507-5.215]) and administration of IV TCZ (OR 2.327 [1.244-4.353]) and previous therapies to TCZ (OR 5.062[2.402-10.665]). CONCLUSION: In this series, GCA relapses were observed in 15 % of patients, all of them non-severe. Relapses were associated with TCZ therapeutic schedule, such as IV administration, optimization, delayed initiation and previous therapies to TCZ.